The present invention relates to a method of transmitting data, in particular GSM data, between at least two subscribers.
With previous GSM data circuits, in particular for digitized speech data between mobile stations over a communication network, such as the public telephone system (PSTNxe2x80x94public switched telephone network), GSM data in GSM format has been converted by a transcoder to the frame format of the communication network (64 kbit/s PCM for PSTN). With the DMCS 900 mobile wireless system, such transcoding takes place at the base stations or at the A interface in transfer to the PSTN communication network. According to GSM Recommendation 0860/0861, digital speech transmission takes place in TRAU frames (TRAU=transcode and rate adaptor unit), i.e., transmission frames in 16 kbit/s format for FR (full rate) transmission and 8 kbit/s or 16 kbit/s for HR (half rate) transmission.
U.S. Pat. No. 5,091,945 describes that instead of speech data, it is possible to transmit only signal parameters (speech parameters) from which the user data can be reconstructed. EP 0 332 345 describes that speech sampling values can be transmitted in a first data stream and a synchronization model can be transmitted.
International patent application WO 96/23297, which was first published after the priority date of the present application, describes that speech sampling values can be transmitted in a first data stream and identification information and speech parameters for speech-data reconstruction can be transmitted in a second data stream. A substantial disadvantage of the data-transmission method described in WO 96/23297 is that the identification information is permanent. WO 96/23297 corresponds to the European patent application EP 806 032.
According to the present invention, an improvement is achieved in the quality of the data transmitted, specifically GSM speech data, in particular in tandem operation with connections between mobile stations. Due to the simultaneous transmission of samples, e.g., PCM values, and signal parameters for reconstruction of the speech data, this reduces interference when using a transcoder that is not equipped for tandem-free operation with transparent switching of TRAU frames. Even when identification information is transmitted, speech data is still transmitted in the first data stream.
The present invention also improves handover operation. The present invention can be implemented easily in existing systems. Only the speech transcoders of a network are affected. For example, of the 8 bits per speech sample, the two least-significant bits are replaced by a TRAU frame (speech is then reduced to 6 bits) into which identification information and/or synchronization information can be inserted to synchronize the transcoders.
By squeezing the lowest-order bits to zero in transmission of speech samples when establishing a voice connection, the speech quality is improved.
In a handover, i.e., when changing to another transcoder, frame loss can be ascertained rapidly, and it is possible to prevent invalid frames from being processed further as TRAU frames. A transcoder that is not equipped for tandem-free operation can process speech samples during a synchronization operation. If a reduced idle pattern is sent instead of samples after synchronization is concluded, the speech decoder of the transcoder can be switched off or used for other purposes, e.g., for full rate/half rate code conversion in the downlink path.